This seems a good one for someone to sink their teeth into. Any takers? OK, here's the problem: Our Engineers are working on a drives application that uses Cisco Switches. I've attached a .tif file that contains 3 pages - a System Diagram, a Logical Diagram and a diagram showing some troubleshooting steps they've done. System Diagram: They've got a PLC with two 10BT copper ethernet cards in it. One card does 10BT half-duplex TX data flow of 1700 UDP Frames per second. The other does RX same type & rate. (There are no collisions.) These are plugged into one of two Cisco 2900 XL fiber switches (the 2 switches are connected via 100FX) through a copper expansion card. The 2 2900XLs are connected via 100FX (autoneg, I believe, for full/half dup) to 17 Digital Drives. Each of the 100FX lines to the drives has data flow of 100 UDP Frames per second for both TX & RX. Logical Diagram: Logically, the PLC talks to each Drive at 100 UDP Frames per second. There's a Protection Timer on each Drive that tracks the max time between UDPs Received. Typical time = 10 milliseconds. Worst case = 30 millisecs. Fail = 1 second +/- 10 milliseconds. Problem: On 1-2 drives randomly every 24 hours we see a loss of PLC to Drive UDP Frames for *exactly* 1 second +/- 10 milliseconds BUT *no* loss of Drive to PLC UDP Frames. Troubleshooting steps thus far: 1. Each drive has 2 LEDs, one for TX Activity and one for RX Activity. They connected the LEDs (or maybe the thing that lights the LEDs?) to a Recorder to monitor the TX & RX. They saw a loss of RX Activity for 1 second and *no* loss of TX Activity. 2. Via diagnostic configuration of the switch, they took Drives #1-4 and sent the Frames they were sending to a Sniffer (which was using custom software that the Engineer in charge wrote). This test was unsuccessful because *even though it's a switch & not a hub* they didn't get data unique to Drives #1-4, but ocassionally got stuff from the other Drives, too. This really stumped them because a switch is not shared media, but switched (duh) & they don't think that they should see anything other than traffic to/from Drives 1-4 since they're each on their own port. Next Troubleshooting steps: 1. Set the Switch to Drive 100FX runs to Full Duplex only 2. Put a 10BT hub between the Switch and the PLC & sniff to see what traffic is happening there. Questions: 1. Why were they seeing stuff from other ports when they did Troubleshooting step #2? 2. Is this a known issue with these switches? It's intermittent in that it doesn't happen with any seeming regularity to any one port, but happens to each port at random intervals. Some ports never do it. Some do it more often than others. Haven't seen any pattern to it. Caveat: They have NOT ruled out the possibility that it's a problem with the PLC. Next Troubleshooting Step #2 is designed to determine if the error is in the switch or the PLC. ie. is the PLC sending the data or is *it* stopping for 1 sec randomly - maybe due to buffer overrun? Any help that anyone can provide - whether it be additional troubleshooting steps or KB articles for the Cisco switches etc. - would be greatly appreciated! Kris Jacobs CCNA, Network+ [GroupStudy.com removed an attachment of type image/tiff which had a name of Drives.tif] Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=2530&t=2530 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
